Journal of Neurology, Neurosurgery, and Psychiatry, 1980, 43, 941-947

Clinical and prognostic correlates of EEG inopen-heart surgery patientsK A SOTANIEMI

From the Department of Neurology, University of Oulu, Oulu, Finland

SUMMARY Sixty-five patients undergoing cardiac valve replacement were followed for one yearby electroencephalography (EEG). Occurrence of delta or sharp wave disturbances or low frequencyof dominant activity before operation was found to have prognostic significance. The degree ofEEG change after operation correlated with clinical signs of cerebral involvement, and predictedthe later course.

The risk of cerebral disorders arising duringopen-heart surgery has been established in anumber of clinical and experimental studies.1-4Extracorporeal circulation may expose the cen-tral nervous system (CNS) to disturbances inblood flow,5 and metabolic changes probably dueto decreased oxygen availability have been re-ported.5 6 However, the causes of cerebral com-plications still are controversial and a singledeterminant can rarely be identified.7-9 Despitetechnical improvements, cerebral disorders con-tinue to occur'0-12 suggesting a need to obtainmore objective and detailed information of themethods used for examination of the CNS. It isevident that clinical investigation usually revealsonly the most severe complications; thereforemore accurate indicators of brain damage arerequired.The electroencephalogram (EEG) is useful

for monitoring cerebral circulation during opera-tion.'3 14 Development of the cerebral functionmonitor (CFM),15 16 has made it possible todetect untoward events, but the CFM has notgained unreserved acceptance because it is costlyand offers only a rough estimate of the quantita-tive EEG. Although more sensitive methods forEEG-monitoring during surgery have been pro-posed,'7 conventional EEG recorded before andafter operation remains the most practical toolfor evaluation of the CNS effects of surgery.The present study was undertaken to assess

the clinical correlates and the prognostic valueof the conventional EEG in open-heart surgery.

Address for reprint requests: Kyosti Sotaniemi, MD, Department ofNeurology, University of Oulu, 90220 Oulu 22, Finland.Accepted 14 April 1980

Patients and methods

Patients and operative procedures Sixty-five con-secutive patients undergoing cardiac valve replace-ment surgery were investigated. There were 22women and 43 men aged from 15 to 65 years, themean age being 43-0 + 10-4 years.

Aortic valve replacement was done in 44 cases,mitral in 16 cases and both of the valves wererepaired in five cases. The prostheses were of theBjork-Shiley type. All the patients were operatedupon with moderate hypothermia (oesophagealtemperature 30-32°C). A Rygg-Kyvsgard bubbleoxygenator was used and extracorporeal circulationwas carried out with moderate haemodilution andnon-pulsatile flow. The patients were divided intotwo groups according to the clinical findings at thefirst follow up: Group NC (no cerebral complica-tions, N=33) and Group CC (cerebral complica-tions present, N=28). The groups were similar forage, sex and the main cardiological diagnoses.Design of the study The protocol of the study ispresented in fig 1. Every patient underwent twopreoperative (coded PRE-OP) neurological and EEGinvestigations on the fifth and second day beforeoperation. The investigations were repeated fivetimes (10 days, two months, five months, eightmonths and one year after operation) during thefollow-up period of one year. The examinationsafter operation are coded using the ordinal of theinvestigation (from I to V) together with the abbre-viation FU (follow-up): For example the first post-operative EEG is coded I FU EEG.Neurological evaluation A full neurological investi-gation was performed immediately before or afterthe EEG recordings. Signs present before operationwere not listed as new findings after operation.EEG The EEGs were recorded using a 16-chan-nel machine (Mingograph EEG 16, Elema-Schonan-der) under standard conditions (time constant 0*3 s,

941

copyright. on July 21, 2022 by guest. P

rotected byhttp://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.43.10.941 on 1 O

ctober 1980. Dow

nloaded from

K A Sotaniemi

Code Pre-op+ operation

days5 2 £Invest igation.. * *

timeNE NEEEG EEG

= Pre-op EEGNumber of 65 65patients - 65

Preoperat iveinvestigations

IFU

10

IIFU IIIFU IVFU VFU

months2 5 8 12

i * * * Fig 1 The investigation protocol. TheNE NE NE NE NE codes and timings of the performedEEG EEG EEG EEG EEG investigations are indicated. (PRE-OP =

preoperative; FU=follow-up; post-61 59 58 57 44 operative; NE = neurological evaluation.)t t t t.Postoperative investigations

high pass filter from 70 c/s, amplification 100 uV= 1 cm or 50 AV= 1 cm, paper speed 3 cm/s: patientawake, semi-reclining position). Silver electrodes wereplaced on the scalp according to the international 10-20 system.'8 Four standard settings were used. Therecording time was at least 30 min in all the in-vestigations. Hyperventilation and photic stimulationwere used as activation procedures in all but thefirst follow-up examination.The BEG interpretation was carried out using

conventional methods of evaluation.19-21 The classi-fication of normality was done using the followinggradation: (1) normal; (2) slightly abnormal: domi-nant activity within normal ranges (or slowing notmore than 1 c/s when compared with a previousrecording of the same individual during the follow-up period), short episodes of theta or delta activityseen in some few occasions: (3) moderately abnor-mal: slow dominant activity (<8 c/s) or slowingwith >1 c/s in the same individual; presence of in-termittent delta episodes: (4) severely abnormal: noalpha activity, abundant or continuous delta activity.The mean value of the two recordings before

operation was considered as the basic preoperativestate (PRE-OP EEG) with which the EEGs afteroperation were compared. The two recordings beforeoperation were found to differ from each other inonly four cases: in these cases classification wasdone according to the more abnormal of the record-ings. The frequency of the dominant activity wascounted manually, and the mean value of at leastfour representative activity periods was used.

Student's t test was used in the statistical analyses(the test for paired samples in calculating the differ-ence between preoperative and postoperative results;the test for independent means in comparing thepatient groups with each other).

Results

Clinical aspects Neurological signs of CNS dis-orders after operation were detected in 31patients. Three patients died within eight daysof operation. Brain damage was the cause ofdeath in two patients and one further patientdied of other causes. Symptoms of a hemiparesiswere found in 19 cases, and the remaining

patients exhibit either aphasia, confusion, brainstem or cerebellar signs. These symptoms wereslight and usually reversible. Residual signs werestill present in five cases one year after opera-tion. (The clinical findings in a larger samplefrom which the present patients form a part havebeen reported earlier.12)

CLINICAL CORRELATES OF EEGGeneral EEG evaluation Follow-up of the in-cidences of abnormal EEG in the NC and CCgroups is presented in fig 2. The groups weresimilar before operation, the incidences ofabnormal EEG being 47% (NC) and 52% (CC).By contrast, considerable differences were seenafter operation: an abnormal EEG was seen in67% (22 cases out of 33) in the NC group andin 97% (27 cases out of 28) in the CC group atfirst follow up.The EEG changes improved in both groups but

the CC cases took longer to recover than the NCpatients: the EEG restored to its PRE-OP stateover two months in the NC group, but it tookup to five months in the other group. One yearafter surgery the incidence of an abnormal EEG

0D100j 90E 80o 7060504030

c 20S 10

S.:IPre-op 4

cperatior

D-I NCgroups CCgroup

7

7

/IFU IIFU IIIFU IVFU VFU

I investigation time -

Fig 2 Follow-up of the incidence of abnormal EEGin the postoperatively non-complicated (NC) andcerebrally complicated (CC) patient groups.

942

1-1 -1

copyright. on July 21, 2022 by guest. P

rotected byhttp://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.43.10.941 on 1 O

ctober 1980. Dow

nloaded from

Clinical and prognostic correlates of EEG in open-heart surgery patients

decreased to 29% in the NC group and to 20%in the CC group.EEG impairment after operation was evident

in 39% of the NC cases and in 75% of the CCpatients. Deterioration of a previously normalEEG after operation occurred in 44% and 92%in groups NC and CC respectively, and anincrease of previous abnormalities approved in33% and 60% respectively.Nature of EEG abnormalities The principaltypes of EEG abnormalities are shown in table 1.The greatest difference between the groups afteroperation was seen in delta range abnormalitywhich increased almost sixfold and disappearedmore slowly in the complicated group when com-pared with the non-complicated cases. Generally,the activity patterns were similar in both groupswhile the frequency changes formed the maindifferences.

Delta range activity of diffuse and continuoustype was the main abnormality in 18 cases atthe first follow-up investigation: 14 (78%) ofthe cases displayed neurological signs of hemi-sphere damage. Thus delta activity was seen in12% of the NC cases and in 50% of the CCcases. The clinical groups showed no significantdifferences in the appearance of other forms ofabnormalities (table 1). Two months after sur-gery (II FU) delta range disturbances were not

Table 1 Main abnormal activity in thepostoperatively non-complicated (NC) andcomplicated (CC) patient group

Investigation Number oftime (code) patients

nc cc

PRE-OP 34 31I FU 33 28I1 FU 32 27III FU 32 26IV FU 32 25V FU 24 20

Percentage number of cases withvarious types of abnormal EEG activityDelta Theta Sharp

wavesNC CC NC CC NC CC

5-9 65 38-2 35-5 2-9 9-715*2 39-3 48-5 53-6 3*0 3-60 14-8 43-8 51-9 3-1 7-40 7-7 28-1 30-8 3-1 3 -8O 0 31-3 36-0 3-1 0O 0 29-2 20-0 0 0

seen in the NC group, but it took five months(III FU) for them to disappear in the othergroup.The course of appearance of major forms of

abnormalities is shown in table 2. The mostmarked postoperative (I FU) difference betweenthe groups was the occurrence of bilateral dis-turbances of continuous type which were presentin the majority of the CC cases, but only aquarter of the NC cases had such a distribution.Strictly focal findings were rare in the CC groupand were absent in the NC group. Unilateralinvolvement remained unchanged in the non-complicated cases, and decreased markedly in thecomplicated cases. No particular region showedspecific liability to disturbances and no differ-ences were seen in recovery. Responses to acti-vation procedures were similar in both of theclinical groups.Background activity Follow-up of the dominantfrequency of the background activity is presentedin table 2. The postoperative (I FU) fall in thefrequency was statistically significant in the CCgroup (from 10-1 + 18 c/s to 8-7+2-5 c/s,p<0-01), while the NC group showed only a slightchange in the value (from 10-4 +2-0 to 9-7 + 2-7c/s). The difference between the I FU values ofthe groups was nearly significant (p<0-05). Therecovery was rapid in the NC cases and thePRE-OP state was restored within five monthsafter surgery. By contrast recovery was slow inthe CC cases and the PRE-OP values were notregained during the follow-up period.Hemisphere differences Table 3 shows the fol-low-up of the incidence of abnormalities in thehemispheres. The left side was affected moreoften than the right before operation, but im-pairment after operation involved both hemi-spheres equally. No interhemispheric differenceswere seen in the later course of the NC cases,but in the CC group the left hemisphere re-covered more slowly than the right hemisphere.EEG and clinical correlates Table 4 shows the

Table 2 Follow-up of the frequency of the background rhythm and the distribution of EEG abnormalityInvestigation Frequency of background Number of Percentage of various types of main abnormal activity in the abnormal EEGstime activity cls± SD abnormal EEGs

NC group CC group Focal Diffuse unilateral Diffuse bilateral Diffuse bilateralepisodic continuous

NC CC NC CC NC CC NC CC NC CC

PRE-OP 10-4±2-0 10.0±18* 16 16 6-3 6-3 37-5 43-7 50-0 25-0 6-2 25-0I FU 9-7±2-2 8-7+2.5* 22 27 4-5 7-4 31-8 14-8 45-5 18-5 18-2 59-3II FU 9-9±1-9 9-5±1-6 15 20 0 0 33-3 20 0 46-7 45-0 20-0 35-0III FU 10-4+1 9 9-8±1-5 10 11 0 9.1 30-0 45-4 60-0 18-2 10.0 27-3IV FU 10-2±1-8 9-9±1-4 10 9 0 0 40-0 55-6 50-0 33-3 10-0 11-1V FU 10-5±2-2 9-6±1-4 7 4 0 0 42-9 100-0 57 1 0 0 0

*: p<001

G

943

copyright. on July 21, 2022 by guest. P

rotected byhttp://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.43.10.941 on 1 O

ctober 1980. Dow

nloaded from

K A Sotaniemi

Table 3 Right and left hemisphere EEG findings inthe NC and CC patient groups

Investigation Number of NC group CC grouptime patients (non-complicated (complicated patients)

patients)NC CC Right Left Right Left

hem hem hem hemabnormal abnormal abnormal abnormalEEG EEG EEG EEGN % N % N% N %

PRE-OP 34 31 12 35 3 15 44 1 11 35 5 14 45 2I FU 33 28 18 54 5 19 57 6 23 82 1 26 92 9III FU 32 26 7 21 9 8 25 0 4 15 4 9 34 6V FU 24 20 5 20 8 5 20 8 1 5 0 3 15 0

Table 4 Incidence of clinical signs of cerebraldysfunction in relation to normal and abnormal EEG

Investigation EEG Normal EEG Abnormaltime Total N Clinical signs Total N Clinical signs

of cases present in of cases present inN % N %

PRE-OP 33 0 0 32 4 12 5I FU 12 1 8 3 49 27 55 1II FU 24 3 12 5 35 19 54-3III FU 37 5 13 3 21 7 33 3IV FU 38 1 2 6 19 6 31 6VFU 33 1 30 11 4 364

relationship between the EEG and the clinicalsigns. In general, abnormality of the EEG cor-responded with the clinical evaluation particu-larly when the EEG had been classified as eithernormal or markedly disturbed. Slight EEGabnormalities, however, reflected the clinicalstate less reliably. In the whole sample of EEGand neurological investigations (numbering 344),clinical signs were present in 6%, 31% and 70%,'of the cases exhibiting a normal EEG, a slightlyabnormal EEG and a moderately-severelyabnormal EEG respectively.

Figure 3 presents the follow-up of the inci-dence of normal and abnormal EEG in relationto the presence or absence of clinical signs ofCNS involvement. Before operation all fourpatients displaying cerebral signs also had EEGdisturbances. At the first follow-up 10 daysafter operation) 27 of the 49 cases (55%) withabnormal EEGs had neurological disorders, butonly one of the 12 cases (8 %) showing a normalEEG displayed clinical signs (those of a brainstem lesion).

Analysis of the cases who exhibited hemi-paresis after operation revealed interesting results.Those displaying left-sided signs (N= 15) showedmore marked EEG abnormality than the caseswith right-sided signs (N=4). All except of onethose with left-sided signs showed bilateral EEG

EEG Status

Normal r i NormalL Abnormal

Abnormal [ Abnormal10090 .80

L, 70

tno6a2u500 40

30') 20

10

0 Pre-op t I FU IIFU IIIFU IVFU VFUoperation .investigation time

Fig 3 Relationship between EEG and clinicalsigns of cerebral involvement through the follow-upperiod.

deterioration, which was seen in only twopatients with right-sided signs: in the remainingcases the EEG and clinical findings were com-patible in both groups. Comparison of the hemi-spheres in later months gave remarkable results.In all those with a left-sided hemiparesis, theright hemisphere displayed a normal EEG eightmonths after operation; the left hemisphererecovered more slowly and still showed abnor-malities in one-quarter of the cases up to theend of the study.Prognostic value of the EEG Some findingsbefore operation were found to have prognosticvalue: (1) transient delta episodes were seen ineight PRE-OP EEGs; six of these cases displayedneurological signs at the first follow-up examina-tion; (2) five cases out of seven having sharpwaves in their PRE-OP EEGs showed signs ofhemispheral affection on the same side afteroperation; (3) when a low (<8 5c/s) frequencyof background activity was observed beforeoperation, the incidence of clinical disorders afteroperation was 70% (seven cases out of 10).Using one or more of the above criteria thepostoperative clinical outcome could be pre-dicted in 15 cases out of 19 (79%).The first EEG after surgery had some value in

predicting the long-term clinical outcome. Allfive patients who displayed residual clinical signsone year after operation, as well as one furtherpatient who developed epilepsy a few monthsafter operation, had moderate or severe disturb-ances in the EEG at the first follow-up examina-tion. Thus six out of the 15 cases showing markedEEG impairment after operation had some kindof long-term neurological disturbance.

944

copyright. on July 21, 2022 by guest. P

rotected byhttp://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.43.10.941 on 1 O

ctober 1980. Dow

nloaded from

Clinical and prognostic correlates of EEG in open-heart surgery patients

Discussion

Open-heart operations are high-risk proceduresand any measure before operation that mightindicate a particularly strong liability to braindamage might be of value in selecting patients.Predictive information would be of specialimportance to those working without the aid ofEEG monitoring during surgery which is usefulas a warning of impending disaster.14 16) 1

Unfortunately, clinical and EEG assessmentsprior to operation generally have failed to predictthe after surgery course13 22 although some prog-nostic information seems to be provided fromneuropsychological studies.2' 24

In the present investigation, non-complicatedand complicated patient groups displayed someEEG differences before operation which were

found to have prognostic importance. Pre-onerative occurrence of bilateral and continuousabnormalities, low frequency of dominantactivity, delta range disturbances and sharp waveabnormalities seemed to be associated withclinical complications after operation. By usingthese preoperative EEG criteria cerebral compli-cations could be predicted in 70% of the casesdisplaying these phenomena, but in only 22% ofthe whole sample. The present results lead usto conclude, like Lorenz and Hehrlein25 that theEEG before surgery mainly indicates the severityof the heart disease producing inadequate cere-bral blood flow. However, the fact that it waspossible to find at least some criteria with prog-nostic significance shows that certain EEGdisturbances and susceptibility to clinical dis-orders are inter-related. All the patients in thissample either had normal or only slightlyabnormal EEGs before operation and a morereliable predictive measure might be related todisturbed cerebral circulation and metabolism.EEG surveillance during operations was not

included in the present study but 17 of thesepatients underwent intraoperative quantitativeEEG monitoring.26 Five of these patients dis-played postoperative clinical signs which were

predicted in four cases from the monitoredEEG, in agreement with previous reports of thepredictive value of EEG information duringsurgery."1 14 16

The first EEG after operation was 10 daysafter surgery, by which time slight disturbancesmay have disappeared. The main aim of thisstudy was to investigate long-term EEG andclinical outcomes and their correlations, so thefirst postoperative recording was at a time whenthe patients were able to carry out normal daily

activities, when even slight clinical disorderswould be recognisable. Also it has been reportedthat EEG abnormalities are not necessarily attheir greatest during the very early postoperativeperiod, but may become more pronouncedseveral days later.27

Neurological symptoms after operation arefrequent. Although a frequency below 10% hasbeen reported,28 the present incidence (48%) isto other earlier results suggesting an incidencebetween 30% and 53%.1-3 The strictness of theclinical criteria seems to be one of the majorvariables in the incidence reported.9 12The EEG findings after operation agreed well

with clinical signs when marked EEG deteriora-tion occurred, as has been reported earlier.8 Goodcorrespondence also was observed when the firstEEG after operation showed no change and par-ticularly when the EEG remained normal. Themost significant EEG variables related to clinicalsigns were the occurrence of delta wave abnor-malities, appearance of bilateral continuousdisturbances and slowing of the backgroundrhythm. The findings support earlier observa-tions,8 25 but also emphasise the importance ofthe changes in the background rhythm.

In contrast the clinical groups showed no dif-ferences in the appearance of focal disturbances,in the deterioration rates of the hemispheres andin responses to activation procedures. The presentresults, with the abundance of clinical signs cffocal damage without corresponding focal EEGabnormalities, but with general and bilateral EEGdisturbances, show that the operative conditionsinfluence the brain as a whole. In view of thenature of the EEG changes, not all of the clinicalsigns can be explained by gross embolisation orany similar regionally limited factors. The resultssupport the concept of postoperative cerebraldysfunction resulting from disturbances in cere-bral metabolism.5 6 29 Metabolic disturbances, inturn, may be generated by a number of factors,such as cerebral hypoperfusion, hypoxia, hypo-thermia, microembolisation of air or antifoamparticles and pharmacological effects, all of whichare potentially present during extracorporeal cir-culation.

Interestingly, the hemispheres seemed to havedefferences in clinical manifestations. Analysis ofthe cases displaying postoperative hemipareticsymptoms revealed that the EEG deteriorationwas most commonly bilateral and affected bothhemispheres equally. However, in the majorityof the cases the clinical signs were generated bythe right hemisphere. There was no obviousexplanation for this interhemispheric disparity,

945

copyright. on July 21, 2022 by guest. P

rotected byhttp://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.43.10.941 on 1 O

ctober 1980. Dow

nloaded from

946

but in all these cases the right hemisphere wasthe dominant one. It might be suggested thatsince the hemispheres differ from each other inseveral properties such as in neuropsychological30and neurophysiological3l 32 functions, there mayalso be differences between the dominant and non-dominant hemispheres in tolerance to exceptionalstrains. Also the thresholds in generating clinicalmanifestations may be different.The possible special characteristics of the hemi-

spheres, for instance in responding to unusualconditions such as prolonged extracorporeal cir-culation, have not been investigated.A considerable ability to recovery, even after

longlasting impairment up to 10 days after sur-gery, seems to be strikingly characteristic of tlheEEGs of cardiac valvular surgery patients; thisconfirms previous findings.33 Encouragingly, thepotential disadvantages of both cardiac valvedisease and surgery, that is the negative conse-quences of prolonged circulatory inadequacy andeven occasional operative cerebral damage, seemto be outweighed by the overall benefits and long-term outcome.

This study was supported in part by a grant fromFinnish Foundation for Cardiovascular Research.

References

1 Gilman S. Cerebral disorders after open-heartoperations. N Engl J Med 1965; 272:489-98.

2 Javid H, Tufo HM, Najafi H, Dye WS, HunterJA, Julian OC. Neurologic abnormalities follow-ing open-heart surgery. J Thorac Cardiovasc Surg1969; 58:502-9.

3 Tufo HM, Ostfeld AM, Shekelle R. Centralnervous system dysfunction following open-heartsurgery. JAMA 1979; 212:1333-40.

4 Wright G, Sanderson JM. Brain damage andmortality in dogs following pulsatile and non-pulsatile blood flow in extra-corporeal circulation.Thorax 1971; 27:738-47.

5 Branthwaite MA. Cerebral blood flow and meta-bolism during open-heart surgery. Thorax 1974;29:633-8.

6 Brennan RW, Patterson RH, Kessler J. Cerebralblood flow and metabolism during cardiopul-monary bypass. Neurology 1971; 21:655-72.

7 Aguilar MJ, Gerbode F, Hill J. Neuropathologiccomplications of cardiac surgery. J ThoracCardiovasc Surg 1971; 61:676-85.

8 Witoszka MM, Tamura H, Indeglia R, HopkinsRW, Simeone FA. Electroencephalographicchanges and cerebral complications in open-heartsurgery. J Thorac Cardiovasc Surg 1973; 66:855-64.

9 Mohr JP. Neurological complications of cardiacvalvular disease and cardiac surgery including

K A Sotaniemi

systemic hyptotension. In: Vinken PJ, Bruyn GW,eds. Handbook of Clinical Neurology, vol 38.Amsterdam: Elsevier, 1979: Part I, 143-67.

10 Ross Russell RW, Bharucha N. The recognitionand prevention of border zone cerebral ischaemiaduring cardiac surgery. Q J Med 1978; 47:303-23.

11 Jaedicke-Hollender K, Jaedicke W, SteinkraussG, Barmeyer J, Spillner G. Psychiatrisch-neurologische Befunde vor und nach aorto-coronarer Bypass-Operationen. Nervenarzt 1979;50:92-101.

12 Sotaniemi KA. Brain damage and neurologicaloutcome after open-heart surgery. J NeurolNeurosurg Psychiatry 1980; 43:127-35.

13 Hansotia PL, Myers WO, Ray JF, Greehling C.Sautter RD. Prognostic value of electroencephalo-gram in cardiac surgery. Ann Thorac Surg 1975;19:127-34.

14 Salerno TA, Lince DP, White DN, Lynn RB,Charette EJP. Monitoring of electroencephalo-gram during open-heart surgery. J ThoracCardiovasc Surg 1978; 54:557-63.

15 Schwartz MS, Colvin MP, Prior PF et al. Thecerebral function monitor. Anaesthesia 1973; 28:611-8.

16 Kritikou PE, Branthwaite MA. Significance ofchanges in cerebral electrical activity at theonset of cardiopulmonary bypass. Thorax 1977;32:534-8.

17 SuIg I, Hokkanen E, Saarela E et al. Compu-terised quantitative EEG and blood pressuremonitoring during high risk surgery. 11th WorldCongress of Neurology. Amsterdam: ExcerptaMedica, International Congress Series No. 427,1977. Abstract No. 522, 175-6.

18 Jasper HH. The ten twenty system of the inter-national federation. Electroencephalogr ClinNeurophysiol 1958; 10:370-5.

19 Gibbs FA, Gibbs EL. Atlas of electroencephalo-graphy III. Reading, Massachusetts: Addison-Wesley, 1958.

20 Gibbs FA, Gibbs EL. Atlas of electroencephalo-graphy II. Reading, Massachusetts: Addison-Wesley, 1959.

21 Gibbs FA, Gibbs EL. Atlas of electroencephalo-graphy III. Reading, Massachusetts: Addison-Wesley, 1964.

22 Torres F, Frank GS, Gohen MM, Lillehei CW,Kaspar N. Neurologic and electroencephalo-graphic studies in open-heart surgery. Neurology1959; 9:174-83.

23 Kilpatrick DG, Miller WC, Allan AN, Lee WH.The use of psychological test data to predictopen-heart surgery outcome: a prospective study.Psychosom Med 1975; 37:62-73.

24 Willner AE, Rabiner CJ, Wisoff BG et al.Analogy tests and psychopathology at follow-upafter open-heart surgery. Biol Psychiat 1976; 11:678-96.

25 Lorenz R, Hehrlein F. Electroencephalographicfindings in heart surgery. Minn Med 1970; 53:1069-76.

copyright. on July 21, 2022 by guest. P

rotected byhttp://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.43.10.941 on 1 O

ctober 1980. Dow

nloaded from

Clinical and prognostic correlates of EEG in open-heart surgery patients

26 Saarela E, Sulg J, Arranto J, Hokkanen E,

Sotaniemi K, Hollmen A. Computerised quanti-tative EEG and blood pressure monitoring duringopen-heart surgery. Fifth Congress of Anaes-thesiology, Paris 1978. Excerpta Medica, Inter-national Congress Series 452, Abstract No. 411.

27 Sachdev NS, Carter CC, Swank RL, Blachly Ph.Relationship between post-cardiotomy delirium,clinical neurological changes, and EEG abnor-malities. J Thorac Cardiovasc Surg 1967; 54:557-63.

28 Branthwaite MA. Prevention of neurologicaldamage during open-heart surgery. Thorax 1975;30:258-61.

29 Vise WM, Schuier F, Hossman K-A, Takagi S,Zulch KJ. Cerebral microembolisation. ArchNeurol 1977; 34:660-5.

30 Milner B. Interhemispheric differences and psy-chological processes. Br Med Bull 1971; 27:272-7.

31 Hoovey ZB, Heinemann U, Creutzfeldt OD.Inter-hemispheric "synchrony" of alpha waves.Electroencephalogr Clin Neurophysiol 1972; 32:337-47.

32 Pfurtscheller G, Maresch H, Schuy S. Inter- andintrahemispheric differences in the peak fre-quency of rhythmic activity within the alphaband. Electroencephalogr Clin Neurophysiol1977; 42:77-83.

33 Arfel G, Weiss J, DuBouchet N. EEG findingsduring open heart surgery with extra-corporealcirculation. In: Gastaut H, Meyer JS, eds.Cerebral Anoxia and the Electroencephalogram.Springfield: Charles C Thomas, 1961: 231-49.

947

copyright. on July 21, 2022 by guest. P

rotected byhttp://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.43.10.941 on 1 O

ctober 1980. Dow

nloaded from